Ping Mei

Short-term precipitation forecasting is a core task in meteorological science, aiming to achieve accurate predictions by modeling the spatiotemporal evolution of radar echo sequences, thereby supporting meteorological services and disaster warning systems. However, existing spatiotemporal sequence prediction methods still struggle to disentangle complex spatiotemporal dependencies effectively and fail to capture the nonlinear chaotic characteristics of precipitation systems. This often results in ambiguous predictions, attenuation of echo intensity, and spatial localization errors. To address these challenges, this paper proposes a unified spatiotemporal sequence prediction framework based on spatiotemporal masking, which comprises two stages: self-supervised pre-training and task-oriented fine-tuning. During pre-training, the model learns global structural features of meteorological systems from sparse contexts by randomly masking local spatiotemporal regions of radar images. In the fine-tuning stage, considering the importance of the temporal dimension in short-term precipitation forecasting and the complex long-range dependencies in spatiotemporal evolution of precipitation systems, we design an RNN-based cyclic temporal mask self-encoder model (MAE-RNN) and a transformer-based spatiotemporal attention model (STMT). The former focuses on capturing short-term temporal dynamics, while the latter simultaneously models long-range dependencies across space and time via a self-attention mechanism, thereby avoiding the smoothing effects on high-frequency details that are typical of conventional convolutional or recurrent structures. The experimental results show that STMT improves 3.73% and 2.39% in CSI and HSS key indexes compared with the existing advanced models, and generates radar echo sequences that are closer to the real data in terms of air mass morphology evolution and reflection intensity grading.

Machine learning has achieved significant success in many statistical application scenarios, but has yet to be fully successful in monthly and seasonal predictions. We identified three statistical challenges in climate prediction: instability of statistical models, complexity of feature factors, and the nonlinearity of the relationship between predictors and predictands. These characteristics limit both traditional empirical forecasting and machine learning methods. This paper proposes a novel method called dynamically modeled machine learning to predict monthly temperature anomalies over China. The core idea of dynamic modeling is that the machine learning model is trained using a sliding time window, so that the relationship between predictors and predictands is optimized for a specific and recent period rather than for the entire time span. One hundred thirty indices related to atmospheric and oceanic circulation and other climatic events from the Beijing Climate Center are used as the feature set. After feature engineering, including feature selection and dimensionality reduction, the predictors are generated and input into a regressor. Five machine learning algorithms are employed as regressors one by one: linear regression (LR), ridge regression (RR), random forest (RF), support vector machine (SVM), and gradient boosting decision trees (GBDTs). The method performs reforecasting for 2012–2021 and compares the results with the output of operational climate models from ECMWF, NCEP, and the Beijing Climate Center. Three quantitative evaluation metrics—predictive score (PS), anomaly correlation coefficient (ACC), and anomaly sign agreement rate—were used to assess the prediction performance of each machine learning regressor, the ensemble model, and three dynamic models. The results demonstrate that the method using GBDTs as the regressor achieves the best predictive performance compared to other methods and operational models, with a monthly average PS score of 84, an ACC value of 0.27, and an anomaly sign agreement rate of 74%.

The Yunnan province of China is a typical humid region but with several severe region-wide droughts. Drought indices are generally used to identify and characterize drought events, and then play a key role in drought prediction. Therefore, a novel prediction model was proposed to predict a comprehensive drought indicator (meteorological composite index, MCI) in Yunnan province. This model combined the recurrent neural networks (RNN) based on a gated recurrent neural unit (GRU) and convolutional neural networks (CNN) with optimization using the modified particle swarm optimization (PSO) algorithm. In this model, pre-processed predictor data were input into the GRU module to extract the time features of the sequences. Furthermore, the feature matrices were input into the CNN module to extract the deep local features and the inter-relationship of the predictors. The model was trained and used to predict the monthly MCI drought index of the representative five stations of Yunnan province from 1960 to 2020. The combined model was evaluated by comparison with traditional machine learning models such as the least absolute shrinkage and selection operator (LASSO) and random forest (RF), and the traditional GRU model. The results show significantly improved skills in root mean square error, mean absolute error and Nash–Sutcliffe efficiency coefficient. This novel method was valuable for the monthly drought prediction in Yunnan province and related climate-risk management.